Bayesian analysis of pig growth curves combining pedigree and genomic information

Abstract

We proposed a genome association study for pig growth curves based on Bayesian hierarchical framework considering different sets of SNP markers and pedigree. Additionally, we aimed also to identify possible chromosome regions affecting the growth curve parameters using empirical weight-age data from an outbred F2 (Brazilian Piau vs commercial) pig population. Under the proposed hierarchical approach, individual growth trajectories were modeled by the nonlinear Gompertz function, so that the parameter estimates were considered to be affected by additive polygenic, systematic and SNP markers effects. The model assuming jointly pedigree and SNP markers presented the best fit based on Deviance Information Criterion. Heritability estimates ranged from 0.53 to 0.56 and from 0.55 to 0.57, respectively for the parameters mature weight (a) and maturing rate (k). Additionally, we found high and positive genetic correlation (0.78) between “a” and "k". The percentages of the genetic variances explained by each SNP allowed identifying the most relevant chromosome regions for each phenotype (growth curve parameters). The majority of these regions were closed to QTL regions previously reported for growth traits. However, we identified three relevant SNPs (55840514 bp at SSC17, 55814469 at SSC17 and 76475804 at SSC X) affecting "a" and "k" simultaneously, and three SNPs affecting only "a" (292758 bp at SSC1, 67319 bp at SSC8 and 50290193 bp at SSC17), that are located in regions not previously described as QTL for growth traits in pigs.